Since 2016, ecological water replenishment (EWR) of the Yongding River has been an important project implemented in response to China's policy of developing an ecological civilization. To study the effect of ecological water replenishment on the aquatic plant community, we took the riparian wetland of the Yongding River (Beijing Plain Section) as the study area. First, the periodic water level variation was analyzed and predicted using Morlet wavelet analysis and NARX neural networks. The annual average water level has the first primary period on the 28a time scale, and the water level shows a dryabundant- dry trend on the 28a time series. The maximum annual absolute error of the NARX neural network is only 0.21%. The forecasting model meets the requirements of water level forecasting for the Yongding River, which has proven to play a leading role in flood control work. By constructing the MIKE11 model of a riparian wetland and comparing the changes in water replenishment discharge and water level simulated over 3 years, it is analyzed that when the replenishment water flow is too high, the restoration effect of the aquatic plant community is slowed down. The restoration effect of the aquatic plant community is poor due to low water flow replenishment. Considering the economy of ecological water replenishment, based on the optimal allocation scheme of ecological water replenishment for the restoration of typical aquatic plant communities in riparian wetlands, the replenishment water flow should be regulated within the range of 18.66~21.33m3/s in 2019 so that the water level in the upper, middle, and lower reaches is maintained at 60.9m, 37.7m, and 27.7-27.8m, respectively.